libcpp/internal.h

   1 /* Part of CPP library.
   2    Copyright (C) 1997-2023 Free Software Foundation, Inc.
   3
   4 This program is free software; you can redistribute it and/or modify it
   5 under the terms of the GNU General Public License as published by the
   6 Free Software Foundation; either version 3, or (at your option) any
   7 later version.
   8
   9 This program is distributed in the hope that it will be useful,
  10 but WITHOUT ANY WARRANTY; without even the implied warranty of
  11 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
  12 GNU General Public License for more details.
  13
  14 You should have received a copy of the GNU General Public License
  15 along with this program; see the file COPYING3.  If not see
  16 <http://www.gnu.org/licenses/>.  */
  17
  18 /* This header defines all the internal data structures and functions
  19    that need to be visible across files.  It should not be used outside
  20    cpplib.  */
  21
  22 #ifndef LIBCPP_INTERNAL_H
  23 #define LIBCPP_INTERNAL_H
  24
  25 #include "symtab.h"
  26 #include "cpplib.h"
  27
  28 #if HAVE_ICONV
  29 #include <iconv.h>
  30 #else
  31 #define HAVE_ICONV 0
  32 typedef int iconv_t;  /* dummy */
  33 #endif
  34
  35 #ifdef __cplusplus
  36 extern "C" {
  37 #endif
  38
  39 struct directive;               /* Deliberately incomplete.  */
  40 struct pending_option;
  41 struct op;
  42 struct _cpp_strbuf;
  43
  44 typedef bool (*convert_f) (iconv_t, const unsigned char *, size_t,
  45                            struct _cpp_strbuf *);
  46 struct cset_converter
  47 {
  48   convert_f func;
  49   iconv_t cd;
  50   int width;
  51   const char* from;
  52   const char* to;
  53 };
  54
  55 #define BITS_PER_CPPCHAR_T (CHAR_BIT * sizeof (cppchar_t))
  56
  57 /* Test if a sign is valid within a preprocessing number.  */
  58 #define VALID_SIGN(c, prevc) \
  59   (((c) == '+' || (c) == '-') && \
  60    ((prevc) == 'e' || (prevc) == 'E' \
  61     || (((prevc) == 'p' || (prevc) == 'P') \
  62         && CPP_OPTION (pfile, extended_numbers))))
  63
  64 #define DIGIT_SEP(c) ((c) == '\'' && CPP_OPTION (pfile, digit_separators))
  65
  66 #define CPP_OPTION(PFILE, OPTION) ((PFILE)->opts.OPTION)
  67 #define CPP_BUFFER(PFILE) ((PFILE)->buffer)
  68 #define CPP_BUF_COLUMN(BUF, CUR) ((CUR) - (BUF)->line_base)
  69 #define CPP_BUF_COL(BUF) CPP_BUF_COLUMN(BUF, (BUF)->cur)
  70
  71 #define CPP_INCREMENT_LINE(PFILE, COLS_HINT) do { \
  72     const class line_maps *line_table = PFILE->line_table; \
  73     const struct line_map_ordinary *map = \
  74       LINEMAPS_LAST_ORDINARY_MAP (line_table); \
  75     linenum_type line = SOURCE_LINE (map, line_table->highest_line); \
  76     linemap_line_start (PFILE->line_table, line + 1, COLS_HINT); \
  77   } while (0)
  78
  79 /* Host alignment handling.  */
  80 struct dummy
  81 {
  82   char c;
  83   union
  84   {
  85     double d;
  86     int *p;
  87   } u;
  88 };
  89
  90 #define DEFAULT_ALIGNMENT offsetof (struct dummy, u)
  91 #define CPP_ALIGN2(size, align) (((size) + ((align) - 1)) & ~((align) - 1))
  92 #define CPP_ALIGN(size) CPP_ALIGN2 (size, DEFAULT_ALIGNMENT)
  93
  94 #define _cpp_mark_macro_used(NODE)                                      \
  95   (cpp_user_macro_p (NODE) ? (NODE)->value.macro->used = 1 : 0)
  96
  97 /* A generic memory buffer, and operations on it.  */
  98 typedef struct _cpp_buff _cpp_buff;
  99 struct _cpp_buff
 100 {
 101   struct _cpp_buff *next;
 102   unsigned char *base, *cur, *limit;
 103 };
 104
 105 extern _cpp_buff *_cpp_get_buff (cpp_reader *, size_t);
 106 extern void _cpp_release_buff (cpp_reader *, _cpp_buff *);
 107 extern void _cpp_extend_buff (cpp_reader *, _cpp_buff **, size_t);
 108 extern _cpp_buff *_cpp_append_extend_buff (cpp_reader *, _cpp_buff *, size_t);
 109 extern void _cpp_free_buff (_cpp_buff *);
 110 extern unsigned char *_cpp_aligned_alloc (cpp_reader *, size_t);
 111 extern unsigned char *_cpp_unaligned_alloc (cpp_reader *, size_t);
 112
 113 #define BUFF_ROOM(BUFF) (size_t) ((BUFF)->limit - (BUFF)->cur)
 114 #define BUFF_FRONT(BUFF) ((BUFF)->cur)
 115 #define BUFF_LIMIT(BUFF) ((BUFF)->limit)
 116
 117 /* #include types.  */
 118 enum include_type
 119   {
 120    /* Directive-based including mechanisms.  */
 121    IT_INCLUDE,  /* #include */
 122    IT_INCLUDE_NEXT,  /* #include_next */
 123    IT_IMPORT,   /* #import  */
 124
 125    /* Non-directive including mechanisms.  */
 126    IT_CMDLINE,  /* -include */
 127    IT_DEFAULT,  /* forced header  */
 128    IT_MAIN,     /* main, start on line 1 */
 129    IT_PRE_MAIN,  /* main, but there will be a preamble before line
 130                     1 */
 131
 132    IT_DIRECTIVE_HWM = IT_IMPORT + 1,  /* Directives below this.  */
 133    IT_HEADER_HWM = IT_DEFAULT + 1     /* Header files below this.  */
 134   };
 135
 136 union utoken
 137 {
 138   const cpp_token *token;
 139   const cpp_token **ptoken;
 140 };
 141
 142 /* A "run" of tokens; part of a chain of runs.  */
 143 typedef struct tokenrun tokenrun;
 144 struct tokenrun
 145 {
 146   tokenrun *next, *prev;
 147   cpp_token *base, *limit;
 148 };
 149
 150 /* Accessor macros for struct cpp_context.  */
 151 #define FIRST(c) ((c)->u.iso.first)
 152 #define LAST(c) ((c)->u.iso.last)
 153 #define CUR(c) ((c)->u.trad.cur)
 154 #define RLIMIT(c) ((c)->u.trad.rlimit)
 155
 156 /* This describes some additional data that is added to the macro
 157    token context of type cpp_context, when -ftrack-macro-expansion is
 158    on.  */
 159 typedef struct
 160 {
 161   /* The node of the macro we are referring to.  */
 162   cpp_hashnode *macro_node;
 163   /* This buffer contains an array of virtual locations.  The virtual
 164      location at index 0 is the virtual location of the token at index
 165      0 in the current instance of cpp_context; similarly for all the
 166      other virtual locations.  */
 167   location_t *virt_locs;
 168   /* This is a pointer to the current virtual location.  This is used
 169      to iterate over the virtual locations while we iterate over the
 170      tokens they belong to.  */
 171   location_t *cur_virt_loc;
 172 } macro_context;
 173
 174 /* The kind of tokens carried by a cpp_context.  */
 175 enum context_tokens_kind {
 176   /* This is the value of cpp_context::tokens_kind if u.iso.first
 177      contains an instance of cpp_token **.  */
 178   TOKENS_KIND_INDIRECT,
 179   /* This is the value of cpp_context::tokens_kind if u.iso.first
 180      contains an instance of cpp_token *.  */
 181   TOKENS_KIND_DIRECT,
 182   /* This is the value of cpp_context::tokens_kind when the token
 183      context contains tokens resulting from macro expansion.  In that
 184      case struct cpp_context::macro points to an instance of struct
 185      macro_context.  This is used only when the
 186      -ftrack-macro-expansion flag is on.  */
 187   TOKENS_KIND_EXTENDED
 188 };
 189
 190 typedef struct cpp_context cpp_context;
 191 struct cpp_context
 192 {
 193   /* Doubly-linked list.  */
 194   cpp_context *next, *prev;
 195
 196   union
 197   {
 198     /* For ISO macro expansion.  Contexts other than the base context
 199        are contiguous tokens.  e.g. macro expansions, expanded
 200        argument tokens.  */
 201     struct
 202     {
 203       union utoken first;
 204       union utoken last;
 205     } iso;
 206
 207     /* For traditional macro expansion.  */
 208     struct
 209     {
 210       const unsigned char *cur;
 211       const unsigned char *rlimit;
 212     } trad;
 213   } u;
 214
 215   /* If non-NULL, a buffer used for storage related to this context.
 216      When the context is popped, the buffer is released.  */
 217   _cpp_buff *buff;
 218
 219   /* If tokens_kind is TOKEN_KIND_EXTENDED, then (as we thus are in a
 220      macro context) this is a pointer to an instance of macro_context.
 221      Otherwise if tokens_kind is *not* TOKEN_KIND_EXTENDED, then, if
 222      we are in a macro context, this is a pointer to an instance of
 223      cpp_hashnode, representing the name of the macro this context is
 224      for.  If we are not in a macro context, then this is just NULL.
 225      Note that when tokens_kind is TOKEN_KIND_EXTENDED, the memory
 226      used by the instance of macro_context pointed to by this member
 227      is de-allocated upon de-allocation of the instance of struct
 228      cpp_context.  */
 229   union
 230   {
 231     macro_context *mc;
 232     cpp_hashnode *macro;
 233   } c;
 234
 235   /* This determines the type of tokens held by this context.  */
 236   enum context_tokens_kind tokens_kind;
 237 };
 238
 239 struct lexer_state
 240 {
 241   /* 1 if we're handling a directive.  2 if it's an include-like
 242      directive.  */
 243   unsigned char in_directive;
 244
 245   /* Nonzero if in a directive that will handle padding tokens itself.
 246      #include needs this to avoid problems with computed include and
 247      spacing between tokens.  */
 248   unsigned char directive_wants_padding;
 249
 250   /* True if we are skipping a failed conditional group.  */
 251   unsigned char skipping;
 252
 253   /* Nonzero if in a directive that takes angle-bracketed headers.  */
 254   unsigned char angled_headers;
 255
 256   /* Nonzero if in a #if or #elif directive.  */
 257   unsigned char in_expression;
 258
 259   /* Nonzero to save comments.  Turned off if discard_comments, and in
 260      all directives apart from #define.  */
 261   unsigned char save_comments;
 262
 263   /* Nonzero if lexing __VA_ARGS__ and __VA_OPT__ are valid.  */
 264   unsigned char va_args_ok;
 265
 266   /* Nonzero if lexing poisoned identifiers is valid.  */
 267   unsigned char poisoned_ok;
 268
 269   /* Nonzero to prevent macro expansion.  */
 270   unsigned char prevent_expansion;
 271
 272   /* Nonzero when parsing arguments to a function-like macro.  */
 273   unsigned char parsing_args;
 274
 275   /* Nonzero if prevent_expansion is true only because output is
 276      being discarded.  */
 277   unsigned char discarding_output;
 278
 279   /* Nonzero to skip evaluating part of an expression.  */
 280   unsigned int skip_eval;
 281
 282   /* Nonzero when tokenizing a deferred pragma.  */
 283   unsigned char in_deferred_pragma;
 284
 285   /* Count to token that is a header-name.  */
 286   unsigned char directive_file_token;
 287
 288   /* Nonzero if the deferred pragma being handled allows macro expansion.  */
 289   unsigned char pragma_allow_expansion;
 290
 291   /* Nonzero if _Pragma should not be interpreted.  */
 292   unsigned char ignore__Pragma;
 293 };
 294
 295 /* Special nodes - identifiers with predefined significance.  */
 296 struct spec_nodes
 297 {
 298   cpp_hashnode *n_defined;              /* defined operator */
 299   cpp_hashnode *n_true;                 /* C++ keyword true */
 300   cpp_hashnode *n_false;                /* C++ keyword false */
 301   cpp_hashnode *n__VA_ARGS__;           /* C99 vararg macros */
 302   cpp_hashnode *n__VA_OPT__;            /* C++ vararg macros */
 303
 304   enum {M_EXPORT, M_MODULE, M_IMPORT, M__IMPORT, M_HWM};
 305
 306   /* C++20 modules, only set when module_directives is in effect.
 307      incoming variants [0], outgoing ones [1] */
 308   cpp_hashnode *n_modules[M_HWM][2];
 309 };
 310
 311 typedef struct _cpp_line_note _cpp_line_note;
 312 struct _cpp_line_note
 313 {
 314   /* Location in the clean line the note refers to.  */
 315   const unsigned char *pos;
 316
 317   /* Type of note.  The 9 'from' trigraph characters represent those
 318      trigraphs, '\\' an escaped newline, ' ' an escaped newline with
 319      intervening space, 0 represents a note that has already been handled,
 320      and anything else is invalid.  */
 321   unsigned int type;
 322 };
 323
 324 /* Represents the contents of a file cpplib has read in.  */
 325 struct cpp_buffer
 326 {
 327   const unsigned char *cur;        /* Current location.  */
 328   const unsigned char *line_base;  /* Start of current physical line.  */
 329   const unsigned char *next_line;  /* Start of to-be-cleaned logical line.  */
 330
 331   const unsigned char *buf;        /* Entire character buffer.  */
 332   const unsigned char *rlimit;     /* Writable byte at end of file.  */
 333   const unsigned char *to_free;    /* Pointer that should be freed when
 334                                       popping the buffer.  */
 335
 336   _cpp_line_note *notes;           /* Array of notes.  */
 337   unsigned int cur_note;           /* Next note to process.  */
 338   unsigned int notes_used;         /* Number of notes.  */
 339   unsigned int notes_cap;          /* Size of allocated array.  */
 340
 341   struct cpp_buffer *prev;
 342
 343   /* Pointer into the file table; non-NULL if this is a file buffer.
 344      Used for include_next and to record control macros.  */
 345   struct _cpp_file *file;
 346
 347   /* Saved value of __TIMESTAMP__ macro - date and time of last modification
 348      of the assotiated file.  */
 349   const unsigned char *timestamp;
 350
 351   /* Value of if_stack at start of this file.
 352      Used to prohibit unmatched #endif (etc) in an include file.  */
 353   struct if_stack *if_stack;
 354
 355   /* True if we need to get the next clean line.  */
 356   bool need_line : 1;
 357
 358   /* True if we have already warned about C++ comments in this file.
 359      The warning happens only for C89 extended mode with -pedantic on,
 360      or for -Wtraditional, and only once per file (otherwise it would
 361      be far too noisy).  */
 362   bool warned_cplusplus_comments : 1;
 363
 364   /* True if we don't process trigraphs and escaped newlines.  True
 365      for preprocessed input, command line directives, and _Pragma
 366      buffers.  */
 367   bool from_stage3 : 1;
 368
 369   /* At EOF, a buffer is automatically popped.  If RETURN_AT_EOF is
 370      true, a CPP_EOF token is then returned.  Otherwise, the next
 371      token from the enclosing buffer is returned.  */
 372   bool return_at_eof : 1;
 373
 374   /* One for a system header, two for a C system header file that therefore
 375      needs to be extern "C" protected in C++, and zero otherwise.  */
 376   unsigned char sysp;
 377
 378   /* The directory of the this buffer's file.  Its NAME member is not
 379      allocated, so we don't need to worry about freeing it.  */
 380   struct cpp_dir dir;
 381
 382   /* Descriptor for converting from the input character set to the
 383      source character set.  */
 384   struct cset_converter input_cset_desc;
 385 };
 386
 387 /* The list of saved macros by push_macro pragma.  */
 388 struct def_pragma_macro {
 389   /* Chain element to previous saved macro.  */
 390   struct def_pragma_macro *next;
 391   /* Name of the macro.  */
 392   char *name;
 393   /* The stored macro content.  */
 394   unsigned char *definition;
 395
 396   /* Definition line number.  */
 397   location_t line;
 398   /* If macro defined in system header.  */
 399   unsigned int syshdr   : 1;
 400   /* Nonzero if it has been expanded or had its existence tested.  */
 401   unsigned int used     : 1;
 402
 403   /* Mark if we save an undefined macro.  */
 404   unsigned int is_undef : 1;
 405   /* Nonzero if it was a builtin macro.  */
 406   unsigned int is_builtin : 1;
 407 };
 408
 409 /* A cpp_reader encapsulates the "state" of a pre-processor run.
 410    Applying cpp_get_token repeatedly yields a stream of pre-processor
 411    tokens.  Usually, there is only one cpp_reader object active.  */
 412 struct cpp_reader
 413 {
 414   /* Top of buffer stack.  */
 415   cpp_buffer *buffer;
 416
 417   /* Overlaid buffer (can be different after processing #include).  */
 418   cpp_buffer *overlaid_buffer;
 419
 420   /* Lexer state.  */
 421   struct lexer_state state;
 422
 423   /* Source line tracking.  */
 424   class line_maps *line_table;
 425
 426   /* The line of the '#' of the current directive.  */
 427   location_t directive_line;
 428
 429   /* Memory buffers.  */
 430   _cpp_buff *a_buff;            /* Aligned permanent storage.  */
 431   _cpp_buff *u_buff;            /* Unaligned permanent storage.  */
 432   _cpp_buff *free_buffs;        /* Free buffer chain.  */
 433
 434   /* Context stack.  */
 435   struct cpp_context base_context;
 436   struct cpp_context *context;
 437
 438   /* If in_directive, the directive if known.  */
 439   const struct directive *directive;
 440
 441   /* Token generated while handling a directive, if any. */
 442   cpp_token directive_result;
 443
 444   /* When expanding a macro at top-level, this is the location of the
 445      macro invocation.  */
 446   location_t invocation_location;
 447
 448   /* This is the node representing the macro being expanded at
 449      top-level.  The value of this data member is valid iff
 450      cpp_in_macro_expansion_p() returns TRUE.  */
 451   cpp_hashnode *top_most_macro_node;
 452
 453   /* Nonzero if we are about to expand a macro.  Note that if we are
 454      really expanding a macro, the function macro_of_context returns
 455      the macro being expanded and this flag is set to false.  Client
 456      code should use the function cpp_in_macro_expansion_p to know if we
 457      are either about to expand a macro, or are actually expanding
 458      one.  */
 459   bool about_to_expand_macro_p;
 460
 461   /* Search paths for include files.  */
 462   struct cpp_dir *quote_include;        /* "" */
 463   struct cpp_dir *bracket_include;      /* <> */
 464   struct cpp_dir no_search_path;        /* No path.  */
 465
 466   /* Chain of all hashed _cpp_file instances.  */
 467   struct _cpp_file *all_files;
 468
 469   struct _cpp_file *main_file;
 470
 471   /* File and directory hash table.  */
 472   struct htab *file_hash;
 473   struct htab *dir_hash;
 474   struct file_hash_entry_pool *file_hash_entries;
 475
 476   /* Negative path lookup hash table.  */
 477   struct htab *nonexistent_file_hash;
 478   struct obstack nonexistent_file_ob;
 479
 480   /* Nonzero means don't look for #include "foo" the source-file
 481      directory.  */
 482   bool quote_ignores_source_dir;
 483
 484   /* Nonzero if any file has contained #pragma once or #import has
 485      been used.  */
 486   bool seen_once_only;
 487
 488   /* Multiple include optimization.  */
 489   const cpp_hashnode *mi_cmacro;
 490   const cpp_hashnode *mi_ind_cmacro;
 491   bool mi_valid;
 492
 493   /* Lexing.  */
 494   cpp_token *cur_token;
 495   tokenrun base_run, *cur_run;
 496   unsigned int lookaheads;
 497
 498   /* Nonzero prevents the lexer from re-using the token runs.  */
 499   unsigned int keep_tokens;
 500
 501   /* Buffer to hold macro definition string.  */
 502   unsigned char *macro_buffer;
 503   unsigned int macro_buffer_len;
 504
 505   /* Descriptor for converting from the source character set to the
 506      execution character set.  */
 507   struct cset_converter narrow_cset_desc;
 508
 509   /* Descriptor for converting from the source character set to the
 510      UTF-8 execution character set.  */
 511   struct cset_converter utf8_cset_desc;
 512
 513   /* Descriptor for converting from the source character set to the
 514      UTF-16 execution character set.  */
 515   struct cset_converter char16_cset_desc;
 516
 517   /* Descriptor for converting from the source character set to the
 518      UTF-32 execution character set.  */
 519   struct cset_converter char32_cset_desc;
 520
 521   /* Descriptor for converting from the source character set to the
 522      wide execution character set.  */
 523   struct cset_converter wide_cset_desc;
 524
 525   /* Date and time text.  Calculated together if either is requested.  */
 526   const unsigned char *date;
 527   const unsigned char *time;
 528
 529   /* Time stamp, set idempotently lazily.  */
 530   time_t time_stamp;
 531   int time_stamp_kind; /* Or errno.  */
 532
 533   /* A token forcing paste avoidance, and one demarking macro arguments.  */
 534   cpp_token avoid_paste;
 535   cpp_token endarg;
 536
 537   /* Opaque handle to the dependencies of mkdeps.cc.  */
 538   class mkdeps *deps;
 539
 540   /* Obstack holding all macro hash nodes.  This never shrinks.
 541      See identifiers.cc */
 542   struct obstack hash_ob;
 543
 544   /* Obstack holding buffer and conditional structures.  This is a
 545      real stack.  See directives.cc.  */
 546   struct obstack buffer_ob;
 547
 548   /* Pragma table - dynamic, because a library user can add to the
 549      list of recognized pragmas.  */
 550   struct pragma_entry *pragmas;
 551
 552   /* Call backs to cpplib client.  */
 553   struct cpp_callbacks cb;
 554
 555   /* Identifier hash table.  */
 556   struct ht *hash_table;
 557
 558   /* Expression parser stack.  */
 559   struct op *op_stack, *op_limit;
 560
 561   /* User visible options.  */
 562   struct cpp_options opts;
 563
 564   /* Special nodes - identifiers with predefined significance to the
 565      preprocessor.  */
 566   struct spec_nodes spec_nodes;
 567
 568   /* Whether cpplib owns the hashtable.  */
 569   bool our_hashtable;
 570
 571   /* Traditional preprocessing output buffer (a logical line).  */
 572   struct
 573   {
 574     unsigned char *base;
 575     unsigned char *limit;
 576     unsigned char *cur;
 577     location_t first_line;
 578   } out;
 579
 580   /* Used for buffer overlays by traditional.cc.  */
 581   const unsigned char *saved_cur, *saved_rlimit, *saved_line_base;
 582
 583   /* A saved list of the defined macros, for dependency checking
 584      of precompiled headers.  */
 585   struct cpp_savedstate *savedstate;
 586
 587   /* Next value of __COUNTER__ macro. */
 588   unsigned int counter;
 589
 590   /* Table of comments, when state.save_comments is true.  */
 591   cpp_comment_table comments;
 592
 593   /* List of saved macros by push_macro.  */
 594   struct def_pragma_macro *pushed_macros;
 595
 596   /* If non-zero, the lexer will use this location for the next token
 597      instead of getting a location from the linemap.  */
 598   location_t forced_token_location;
 599
 600   /* Location identifying the main source file -- intended to be line
 601      zero of said file.  */
 602   location_t main_loc;
 603
 604   /* Returns true iff we should warn about UTF-8 bidirectional control
 605      characters.  */
 606   bool warn_bidi_p () const
 607   {
 608     return (CPP_OPTION (this, cpp_warn_bidirectional)
 609             & (bidirectional_unpaired|bidirectional_any));
 610   }
 611 };
 612
 613 /* Character classes.  Based on the more primitive macros in safe-ctype.h.
 614    If the definition of `numchar' looks odd to you, please look up the
 615    definition of a pp-number in the C standard [section 6.4.8 of C99].
 616
 617    In the unlikely event that characters other than \r and \n enter
 618    the set is_vspace, the macro handle_newline() in lex.cc must be
 619    updated.  */
 620 #define _dollar_ok(x)   ((x) == '$' && CPP_OPTION (pfile, dollars_in_ident))
 621
 622 #define is_idchar(x)    (ISIDNUM(x) || _dollar_ok(x))
 623 #define is_numchar(x)   ISIDNUM(x)
 624 #define is_idstart(x)   (ISIDST(x) || _dollar_ok(x))
 625 #define is_numstart(x)  ISDIGIT(x)
 626 #define is_hspace(x)    ISBLANK(x)
 627 #define is_vspace(x)    IS_VSPACE(x)
 628 #define is_nvspace(x)   IS_NVSPACE(x)
 629 #define is_space(x)     IS_SPACE_OR_NUL(x)
 630
 631 #define SEEN_EOL() (pfile->cur_token[-1].type == CPP_EOF)
 632
 633 /* This table is constant if it can be initialized at compile time,
 634    which is the case if cpp was compiled with GCC >=2.7, or another
 635    compiler that supports C99.  */
 636 #if HAVE_DESIGNATED_INITIALIZERS
 637 extern const unsigned char _cpp_trigraph_map[UCHAR_MAX + 1];
 638 #else
 639 extern unsigned char _cpp_trigraph_map[UCHAR_MAX + 1];
 640 #endif
 641
 642 #if !defined (HAVE_UCHAR) && !defined (IN_GCC)
 643 typedef unsigned char uchar;
 644 #endif
 645
 646 #define UC (const uchar *)  /* Intended use: UC"string" */
 647
 648 /* Accessors.  */
 649
 650 inline int
 651 _cpp_in_system_header (cpp_reader *pfile)
 652 {
 653   return pfile->buffer ? pfile->buffer->sysp : 0;
 654 }
 655 #define CPP_PEDANTIC(PF) CPP_OPTION (PF, cpp_pedantic)
 656 #define CPP_WTRADITIONAL(PF) CPP_OPTION (PF, cpp_warn_traditional)
 657
 658 /* Return true if we're in the main file (unless it's considered to be
 659    an include file in its own right.  */
 660 inline int
 661 _cpp_in_main_source_file (cpp_reader *pfile)
 662 {
 663   return (!CPP_OPTION (pfile, main_search)
 664           && pfile->buffer->file == pfile->main_file);
 665 }
 666
 667 /* True if NODE is a macro for the purposes of ifdef, defined etc.  */
 668 inline bool _cpp_defined_macro_p (cpp_hashnode *node)
 669 {
 670   /* Do not treat conditional macros as being defined.  This is due to
 671      the powerpc port using conditional macros for 'vector', 'bool',
 672      and 'pixel' to act as conditional keywords.  This messes up tests
 673      like #ifndef bool.  */
 674   return cpp_macro_p (node) && !(node->flags & NODE_CONDITIONAL);
 675 }
 676
 677 /* In macro.cc */
 678 extern bool _cpp_notify_macro_use (cpp_reader *pfile, cpp_hashnode *node,
 679                                    location_t);
 680 inline bool _cpp_maybe_notify_macro_use (cpp_reader *pfile, cpp_hashnode *node,
 681                                          location_t loc)
 682 {
 683   if (!(node->flags & NODE_USED))
 684     return _cpp_notify_macro_use (pfile, node, loc);
 685   return true;
 686 }
 687 extern cpp_macro *_cpp_new_macro (cpp_reader *, cpp_macro_kind, void *);
 688 extern void _cpp_free_definition (cpp_hashnode *);
 689 extern bool _cpp_create_definition (cpp_reader *, cpp_hashnode *, location_t);
 690 extern void _cpp_pop_context (cpp_reader *);
 691 extern void _cpp_push_text_context (cpp_reader *, cpp_hashnode *,
 692                                     const unsigned char *, size_t);
 693 extern bool _cpp_save_parameter (cpp_reader *, unsigned, cpp_hashnode *,
 694                                  cpp_hashnode *);
 695 extern void _cpp_unsave_parameters (cpp_reader *, unsigned);
 696 extern bool _cpp_arguments_ok (cpp_reader *, cpp_macro *, const cpp_hashnode *,
 697                                unsigned int);
 698 extern const unsigned char *_cpp_builtin_macro_text (cpp_reader *,
 699                                                      cpp_hashnode *,
 700                                                      location_t = 0);
 701 extern int _cpp_warn_if_unused_macro (cpp_reader *, cpp_hashnode *, void *);
 702 extern void _cpp_push_token_context (cpp_reader *, cpp_hashnode *,
 703                                      const cpp_token *, unsigned int);
 704 extern void _cpp_backup_tokens_direct (cpp_reader *, unsigned int);
 705
 706 /* In identifiers.cc */
 707 extern void _cpp_init_hashtable (cpp_reader *, cpp_hash_table *);
 708 extern void _cpp_destroy_hashtable (cpp_reader *);
 709
 710 /* In files.cc */
 711 enum _cpp_find_file_kind
 712   { _cpp_FFK_NORMAL, _cpp_FFK_FAKE, _cpp_FFK_PRE_INCLUDE, _cpp_FFK_HAS_INCLUDE };
 713 extern _cpp_file *_cpp_find_file (cpp_reader *, const char *, cpp_dir *,
 714                                   int angle, _cpp_find_file_kind, location_t);
 715 extern bool _cpp_find_failed (_cpp_file *);
 716 extern void _cpp_mark_file_once_only (cpp_reader *, struct _cpp_file *);
 717 extern const char *_cpp_find_header_unit (cpp_reader *, const char *file,
 718                                           bool angle_p,  location_t);
 719 extern void _cpp_fake_include (cpp_reader *, const char *);
 720 extern bool _cpp_stack_file (cpp_reader *, _cpp_file*, include_type, location_t);
 721 extern bool _cpp_stack_include (cpp_reader *, const char *, int,
 722                                 enum include_type, location_t);
 723 extern int _cpp_compare_file_date (cpp_reader *, const char *, int);
 724 extern void _cpp_report_missing_guards (cpp_reader *);
 725 extern void _cpp_init_files (cpp_reader *);
 726 extern void _cpp_cleanup_files (cpp_reader *);
 727 extern void _cpp_pop_file_buffer (cpp_reader *, struct _cpp_file *,
 728                                   const unsigned char *);
 729 extern bool _cpp_save_file_entries (cpp_reader *pfile, FILE *f);
 730 extern bool _cpp_read_file_entries (cpp_reader *, FILE *);
 731 extern const char *_cpp_get_file_name (_cpp_file *);
 732 extern struct stat *_cpp_get_file_stat (_cpp_file *);
 733 extern bool _cpp_has_header (cpp_reader *, const char *, int,
 734                              enum include_type);
 735
 736 /* In expr.cc */
 737 extern bool _cpp_parse_expr (cpp_reader *, bool);
 738 extern struct op *_cpp_expand_op_stack (cpp_reader *);
 739
 740 /* In lex.cc */
 741 extern void _cpp_process_line_notes (cpp_reader *, int);
 742 extern void _cpp_clean_line (cpp_reader *);
 743 extern bool _cpp_get_fresh_line (cpp_reader *);
 744 extern bool _cpp_skip_block_comment (cpp_reader *);
 745 extern cpp_token *_cpp_temp_token (cpp_reader *);
 746 extern const cpp_token *_cpp_lex_token (cpp_reader *);
 747 extern cpp_token *_cpp_lex_direct (cpp_reader *);
 748 extern unsigned char *_cpp_spell_ident_ucns (unsigned char *, cpp_hashnode *);
 749 extern int _cpp_equiv_tokens (const cpp_token *, const cpp_token *);
 750 extern void _cpp_init_tokenrun (tokenrun *, unsigned int);
 751 extern cpp_hashnode *_cpp_lex_identifier (cpp_reader *, const char *);
 752 extern int _cpp_remaining_tokens_num_in_context (cpp_context *);
 753 extern void _cpp_init_lexer (void);
 754 static inline void *_cpp_reserve_room (cpp_reader *pfile, size_t have,
 755                                        size_t extra)
 756 {
 757   if (BUFF_ROOM (pfile->a_buff) < (have + extra))
 758     _cpp_extend_buff (pfile, &pfile->a_buff, extra);
 759   return BUFF_FRONT (pfile->a_buff);
 760 }
 761 extern void *_cpp_commit_buff (cpp_reader *pfile, size_t size);
 762
 763 /* In init.cc.  */
 764 extern void _cpp_maybe_push_include_file (cpp_reader *);
 765 extern const char *cpp_named_operator2name (enum cpp_ttype type);
 766 extern void _cpp_restore_special_builtin (cpp_reader *pfile,
 767                                           struct def_pragma_macro *);
 768
 769 /* In directives.cc */
 770 extern int _cpp_test_assertion (cpp_reader *, unsigned int *);
 771 extern int _cpp_handle_directive (cpp_reader *, bool);
 772 extern void _cpp_define_builtin (cpp_reader *, const char *);
 773 extern char ** _cpp_save_pragma_names (cpp_reader *);
 774 extern void _cpp_restore_pragma_names (cpp_reader *, char **);
 775 extern int _cpp_do__Pragma (cpp_reader *, location_t);
 776 extern void _cpp_init_directives (cpp_reader *);
 777 extern void _cpp_init_internal_pragmas (cpp_reader *);
 778 extern void _cpp_do_file_change (cpp_reader *, enum lc_reason, const char *,
 779                                  linenum_type, unsigned int);
 780 extern void _cpp_pop_buffer (cpp_reader *);
 781 extern char *_cpp_bracket_include (cpp_reader *);
 782
 783 /* In errors.cc  */
 784 extern location_t cpp_diagnostic_get_current_location (cpp_reader *);
 785
 786 /* In traditional.cc.  */
 787 extern bool _cpp_scan_out_logical_line (cpp_reader *, cpp_macro *, bool);
 788 extern bool _cpp_read_logical_line_trad (cpp_reader *);
 789 extern void _cpp_overlay_buffer (cpp_reader *pfile, const unsigned char *,
 790                                  size_t);
 791 extern void _cpp_remove_overlay (cpp_reader *);
 792 extern cpp_macro *_cpp_create_trad_definition (cpp_reader *);
 793 extern bool _cpp_expansions_different_trad (const cpp_macro *,
 794                                             const cpp_macro *);
 795 extern unsigned char *_cpp_copy_replacement_text (const cpp_macro *,
 796                                                   unsigned char *);
 797 extern size_t _cpp_replacement_text_len (const cpp_macro *);
 798
 799 /* In charset.cc.  */
 800
 801 /* The normalization state at this point in the sequence.
 802    It starts initialized to all zeros, and at the end
 803    'level' is the normalization level of the sequence.  */
 804
 805 struct normalize_state
 806 {
 807   /* The previous starter character.  */
 808   cppchar_t previous;
 809   /* The combining class of the previous character (whether or not a
 810      starter).  */
 811   unsigned char prev_class;
 812   /* The lowest normalization level so far.  */
 813   enum cpp_normalize_level level;
 814 };
 815 #define INITIAL_NORMALIZE_STATE { 0, 0, normalized_KC }
 816 #define NORMALIZE_STATE_RESULT(st) ((st)->level)
 817
 818 /* We saw a character C that matches ISIDNUM(), update a
 819    normalize_state appropriately.  */
 820 #define NORMALIZE_STATE_UPDATE_IDNUM(st, c)     \
 821   ((st)->previous = (c), (st)->prev_class = 0)
 822
 823 extern bool _cpp_valid_ucn (cpp_reader *, const unsigned char **,
 824                             const unsigned char *, int,
 825                             struct normalize_state *state,
 826                             cppchar_t *,
 827                             source_range *char_range,
 828                             cpp_string_location_reader *loc_reader);
 829
 830 extern bool _cpp_valid_utf8 (cpp_reader *pfile,
 831                              const uchar **pstr,
 832                              const uchar *limit,
 833                              int identifier_pos,
 834                              struct normalize_state *nst,
 835                              cppchar_t *cp);
 836
 837 extern void _cpp_destroy_iconv (cpp_reader *);
 838 extern unsigned char *_cpp_convert_input (cpp_reader *, const char *,
 839                                           unsigned char *, size_t, size_t,
 840                                           const unsigned char **, off_t *);
 841 extern const char *_cpp_default_encoding (void);
 842 extern cpp_hashnode * _cpp_interpret_identifier (cpp_reader *pfile,
 843                                                  const unsigned char *id,
 844                                                  size_t len);
 845
 846 /* Utility routines and macros.  */
 847 #define DSC(str) (const unsigned char *)str, sizeof str - 1
 848
 849 /* These are inline functions instead of macros so we can get type
 850    checking.  */
 851 static inline int ustrcmp (const unsigned char *, const unsigned char *);
 852 static inline int ustrncmp (const unsigned char *, const unsigned char *,
 853                             size_t);
 854 static inline size_t ustrlen (const unsigned char *);
 855 static inline const unsigned char *uxstrdup (const unsigned char *);
 856 static inline const unsigned char *ustrchr (const unsigned char *, int);
 857 static inline int ufputs (const unsigned char *, FILE *);
 858
 859 /* Use a const char for the second parameter since it is usually a literal.  */
 860 static inline int ustrcspn (const unsigned char *, const char *);
 861
 862 static inline int
 863 ustrcmp (const unsigned char *s1, const unsigned char *s2)
 864 {
 865   return strcmp ((const char *)s1, (const char *)s2);
 866 }
 867
 868 static inline int
 869 ustrncmp (const unsigned char *s1, const unsigned char *s2, size_t n)
 870 {
 871   return strncmp ((const char *)s1, (const char *)s2, n);
 872 }
 873
 874 static inline int
 875 ustrcspn (const unsigned char *s1, const char *s2)
 876 {
 877   return strcspn ((const char *)s1, s2);
 878 }
 879
 880 static inline size_t
 881 ustrlen (const unsigned char *s1)
 882 {
 883   return strlen ((const char *)s1);
 884 }
 885
 886 static inline const unsigned char *
 887 uxstrdup (const unsigned char *s1)
 888 {
 889   return (const unsigned char *) xstrdup ((const char *)s1);
 890 }
 891
 892 static inline const unsigned char *
 893 ustrchr (const unsigned char *s1, int c)
 894 {
 895   return (const unsigned char *) strchr ((const char *)s1, c);
 896 }
 897
 898 static inline int
 899 ufputs (const unsigned char *s, FILE *f)
 900 {
 901   return fputs ((const char *)s, f);
 902 }
 903
 904 /* In line-map.cc.  */
 905
 906 /* Create and return a virtual location for a token that is part of a
 907    macro expansion-list at a macro expansion point.  See the comment
 908    inside struct line_map_macro to see what an expansion-list exactly
 909    is.
 910
 911    A call to this function must come after a call to
 912    linemap_enter_macro.
 913
 914    MAP is the map into which the source location is created.  TOKEN_NO
 915    is the index of the token in the macro replacement-list, starting
 916    at number 0.
 917
 918    ORIG_LOC is the location of the token outside of this macro
 919    expansion.  If the token comes originally from the macro
 920    definition, it is the locus in the macro definition; otherwise it
 921    is a location in the context of the caller of this macro expansion
 922    (which is a virtual location or a source location if the caller is
 923    itself a macro expansion or not).
 924
 925    MACRO_DEFINITION_LOC is the location in the macro definition,
 926    either of the token itself or of a macro parameter that it
 927    replaces.  */
 928 location_t linemap_add_macro_token (const line_map_macro *,
 929                                     unsigned int,
 930                                     location_t,
 931                                     location_t);
 932
 933 /* Return the source line number corresponding to source location
 934    LOCATION.  SET is the line map set LOCATION comes from.  If
 935    LOCATION is the location of token that is part of the
 936    expansion-list of a macro expansion return the line number of the
 937    macro expansion point.  */
 938 int linemap_get_expansion_line (class line_maps *,
 939                                 location_t);
 940
 941 /* Return the path of the file corresponding to source code location
 942    LOCATION.
 943
 944    If LOCATION is the location of a token that is part of the
 945    replacement-list of a macro expansion return the file path of the
 946    macro expansion point.
 947
 948    SET is the line map set LOCATION comes from.  */
 949 const char* linemap_get_expansion_filename (class line_maps *,
 950                                             location_t);
 951
 952 /* A subclass of rich_location for emitting a diagnostic
 953    at the current location of the reader, but flagging
 954    it with set_escape_on_output (true).  */
 955 class encoding_rich_location : public rich_location
 956 {
 957  public:
 958   encoding_rich_location (cpp_reader *pfile)
 959   : rich_location (pfile->line_table,
 960                    cpp_diagnostic_get_current_location (pfile))
 961   {
 962     set_escape_on_output (true);
 963   }
 964
 965   encoding_rich_location (cpp_reader *pfile, location_t loc)
 966   : rich_location (pfile->line_table, loc)
 967   {
 968     set_escape_on_output (true);
 969   }
 970 };
 971
 972 #ifdef __cplusplus
 973 }
 974 #endif
 975
 976 #endif /* ! LIBCPP_INTERNAL_H */